fix(glossary): skip variants inclusion for machinery and checks

Only matching source strings should be included, the variants are useful
for display only.

Fixes #11883

weblate/checks/glossary.py

@@ -25,7 +25,7 @@ def check_single(self, source, target, unit):
         mismatched = set()
         matched = set()
         boundary = r"\b" if unit.translation.language.uses_whitespace() else ""
-        for term in get_glossary_terms(unit):
+        for term in get_glossary_terms(unit, include_variants=False):
             term_source = term.source
             flags = term.all_flags
             expected = term_source if "read-only" in flags else term.target

weblate/checks/same.py

@@ -178,7 +178,7 @@ def should_ignore(self, source, unit) -> bool:
             # Extract untranslatable terms
             terms = [
                 re.escape(term.source)
-                for term in get_glossary_terms(unit)
+                for term in get_glossary_terms(unit, include_variants=False)
                 if "read-only" in term.all_flags
             ]
             if terms:

weblate/glossary/models.py

@@ -71,7 +71,9 @@ def get_glossary_units(project, source_language, target_language):
     )
 
 
-def get_glossary_terms(unit: Unit, *, full: bool = False) -> list[Unit]:
+def get_glossary_terms(
+    unit: Unit, *, full: bool = False, include_variants: bool = True
+) -> list[Unit]:
     """Return list of term pairs for an unit."""
     from weblate.trans.models.component import Component
 
@@ -146,21 +148,23 @@ def get_glossary_terms(unit: Unit, *, full: bool = False) -> list[Unit]:
         # Add variants manually. This could be done by adding filtering on
         # variant__unit__source in the above query, but this slows down the query
         # considerably and variants are rarely used.
-        existing = {match.pk for match in units}
-        variants = set()
-        extra = []
-        for match in units:
-            if not match.variant or match.variant.pk in variants:
-                continue
-            variants.add(match.variant.pk)
-            for child in match.variant.unit_set.filter(
-                translation__language=language
-            ).select_related("source_unit"):
-                if child.pk not in existing:
-                    existing.add(child.pk)
-                    extra.append(child)
-
-        units.extend(extra)
+        if include_variants:
+            existing = {match.pk for match in units}
+            variants = set()
+            extra = []
+
+            for match in units:
+                if not match.variant or match.variant.pk in variants:
+                    continue
+                variants.add(match.variant.pk)
+                for child in match.variant.unit_set.filter(
+                    translation__language=language
+                ).select_related("source_unit"):
+                    if child.pk not in existing:
+                        existing.add(child.pk)
+                        extra.append(child)
+
+            units.extend(extra)
 
         # Order results, this is Python reimplementation of:
         units.sort(key=lambda x: x.glossary_sort_key)

weblate/machinery/microsoft.py

@@ -178,7 +178,7 @@ def format_replacement(
     def get_highlights(self, text, unit):
         result = list(super().get_highlights(text, unit))
 
-        for term in get_glossary_terms(unit):
+        for term in get_glossary_terms(unit, include_variants=False):
             for start, end in term.glossary_positions:
                 glossary_highlight = (start, end, text[start:end], term)
                 handled = False

weblate/machinery/openai.py

@@ -98,7 +98,9 @@ def get_prompt(
         glossary = ""
         if any(units):
             glossary = render_glossary_units_tsv(
-                chain.from_iterable(get_glossary_terms(unit) for unit in units)
+                chain.from_iterable(
+                    get_glossary_terms(unit, include_variants=False) for unit in units
+                )
             )
             if glossary:
                 glossary = GLOSSARY_PROMPT.format(glossary)